The Institution of Railway Signal Engineers (IRSE) celebrated its centenary in 2012. Although a body to represent this specialist branch of rail engineering had been in existence since 1891, formal recognition only arrived with the formation of the IRSE and its first meeting on 25 February 1913 at the Grand Hotel in Birmingham. By that time the Institution had 112 members, mostly senior engineers from the many railway companies but with representatives from the major supply industry also allowed to join.
What better way to celebrate this first meeting than to re-enact it on the exact day 100 years later? The Grand Hotel is closed for refurbishment but the IET premises at Austin Court made for a suitable replacement in Birmingham City Centre. With five gentlemen entering the stage in Edwardian dress (ok – George V was on the throne by then!), these ‘actors’, who are in fact all professional signalling engineers in their own right, gave a rendering of the first meeting’s content. A nice touch to emphasise the now international membership, was to include one gentleman from India and another from The Netherlands.
What did they discuss 100 years ago?
The subject was ‘Signalling and its Connection with the Construction and Management of Railways’ presided over by Mr AT Blackhall, GWR signal engineer and the first IRSE President. The paper was delivered by Mr W Cotterill, chief clerk in the telegraph department of the Midland Railway, on behalf of the paper’s author Mr RJ Insell, chief assistant signal engineer on the GWR, who had been called away to deal with a serious signalling failure. Some things have not changed!
The content of the paper would be as relevant today as it was then. Items included:
» Signal engineers must have extensive knowledge of other engineering departments and rail operations;
» Traffic increases will impact on signal engineers’ duties and responsibilities;
» Signal engineers must influence the track layout in stations and yards and especially the position of junctions;
» Location of signal boxes in relation to pointwork is vitally important since getting it wrong can be very expensive to put right;
» For optimum block working, signal boxes should not be closer than 880 yards;
» When positioning passing loops on single lines, it is important to take account of gradients and over-runs for the safe passage of trains.
Through to the present day
John Francis, a Past President who has a wealth of signalling knowledge acquired from many years in the industry both railway and supply industry based, recounted the major changes that have happened since 1913. Signal boxes were stylish and functional, typically elevated over the track and gradually equipped with kitchen, hot water and sanitation. Styles moved with the architecture of the time: art deco, post war functionalism, modular construction.
Mechanical apparatus gave way to control panels, firstly with switches then later with entrance- exit buttons, and lastly with modern screen-based technology. Signallers can no longer see the trains and the workplace is more akin to an IT office with no opening windows. Modern day dispatchers have VDUs and lots of colour, but wall mounted overview screens are still valued whenever they are provided.
Level crossings have moved from gates (a few still remain) to a multitude of technical solutions for the differing types but are still beset by the eccentric behaviour of the public. A greater onus now rests with the signaller to ensure safe operation and passage of trains at full barrier crossings but road users are responsible for their own safety at the others, often resulting in misuse and tragic accidents.
The traditional train graph in active mode via technology screens has become increasingly useful to manage and regulate train services.
So has this advancement improved the train service offering? Most of it has, but John believes there are contentious points: colour light signals and complex approach controls may have worsened train path utilisation at junctions; platform usage by short trains do not normally allow multiple occupancy that would be easy with mid platform signals or even permissive working. We may have advanced the science but perhaps gone backwards in the art. Signal engineers need to have much greater influence over train operators.
The maintainer’s perspective
The maintainers’ perspective was studied by Peter Halliwell from Network Rail. A dramatic reduction in staff fatalities has happened since 1950; down from 180 to around 5 per annum. Accessing sites has, however, become more difficult – particularly hard for transporting bulky items like point machines.
The Clapham accident and its subsequent report has underpinned the present way of working, with the control of records, a rigid testing regime, single point of responsibility, standardised competence arrangements, fatigue management / control of hours worked, wrong side failure classification and reporting. But is this enough?
Signalling equipment still suffers from reliability problems. Reliance on localised ‘expert’ knowledge is too dominant with consequential difficulties in acquiring hard data. The advancement of remote condition monitoring is taking too long. Yet to be properly considered is the question of who takes responsibility for on board train mounted equipment which is set to become around 50% of the signalling system within a few years. Establishing the necessary inter- relationship between signalling and rolling-stock engineers needs to be addressed with a degree of urgency.
A driver’s view
Has signalling served the driver well? This was considered by Paul Le Vesconte, an ex-driver who became an operations manager. The objective is to arrive safely and on time and in general this is achieved.
However, the many different type of signals and different types of block working remain a source of confusion. The migration from ex- GW ATC (automatic train control) to AWS (automatic warning system) to TPWS (train protection and warning system), but stopping short of full ATP (automatic train protection), has been a welcome improvement but the functionality of AWS has become clouded. Using the system to mark temporary speed restrictions and the approach to an automatic open level crossing (AOCL) can cause misunderstandings. TPWS is good but not that good; it has prevented 80% of SPADs (Signal Passed at Danger). The trial ATP systems on the Great Western and Chiltern main lines were initially unreliable and restrictive on drivers’ techniques. This has improved over time with different driving styles.
The 1999 accident at Ladbroke Grove, just outside Paddington, with the loss of 31 lives was a wake up call on the need for better train protection systems. Playing around with different types of aspects (flashing yellow, flashing green), aimed at improving junction speeds rather than safety, led to at least one significant accident – Colwich in 1986.
LED signals are generally well accepted, being much clearer and more reliable, but please don’t mix different types of signal on the same route. The advent of cab-to- signaller radio systems has been most welcome and the introduction of GSM-R will overcome the shortcomings of earlier systems.
As to the future, the introduction of ERTMS / ETCS on the Cambrian route must be regarded as a useful trial from which many lessons should be learned. The system is not just a change in signalling, it is much more about how a route is operated.
The potential for complexity in the cab is considerable and it must not force complexity into the driving task. The ETCS on board equipment needs to be integrated with on train electronic systems and the display screen (known as the driver machine interface or DMI) must be clearly visible in all lighting conditions.
Some drivers claim that the setup procedures are much too hard, and the system needs to allow the driver to make mistakes, otherwise over cautious driving will result. When failures do occur, as they inevitably will, a robust, degraded mode of operation must be available and the whole system must include provision for possession and staff safety protection.
Whilst control room technology has advanced considerably over the past 25 years, there are several challenges ahead, said Phil Blacker from Atkins which is a major player in the design and provision of signalling systems. Gaining capacity without huge infrastructure costs is one such objective since timetabled train-kilometres are not keeping pace with passenger growth.
Energy efficiency is also important but whilst power for the signalling systems is irrelevant compared to train power demands, the signalling can massively impact on the latter.
Interoperability is a buzz word but it should cost much less than it does and must apply to much more than having trains that can go anywhere. The need for interoperable subsystems and unified interlocking designs is something that the signalling profession is particularly slow to take up and realise the available benefits. The ‘not invented here’ syndrome continues to prevail.
So where should efforts be concentrated in the control system specialism? More is needed to use the latent intelligence of modern computing power, particularly decision support tools and automation. There is a need for standards but these tend to mature with technology and new standards will have to evolve. Communication paths for both vital and non vital applications will need high bandwidth availability, with COTS equipment supplied and structured to internationally-recognised open standards.
Obsolescence is an increasing problem and system lifecycles are shortening. Thus designs should prepare for piecemeal sub system replacement but with the requirement of backward compatibility. The overall objective must be to get better system performance with the ‘Five 9s’ (99.999%) target for availability and reliability. Maybe to achieve this, the signal engineer has to become a systems engineer / specialist engineer with a multi-industry knowledge.
An interview with the future
It was appropriate that a ‘spoof’ interview projecting forward to 2063 should be conducted by two younger members, Padric Dunne and Matt Slade. Looking back on the previous 50 years since 2012, some milestone achievements were recorded:
2019 – Condition monitoring became commonplace with 90% of main lines and 100% of metros equipped. Asset failures in operating hours were virtually eliminated;
2023 – Lineside cables were essentially eliminated at last;
2033 – Social media inputs and data led to dynamic alteration of metro timetables to meet hour by hour demand. External weather conditions allow for control of train braking rates;
2053 – The ‘virtualised cluster’ technology provide for control system hardware. From this technology, it was possible to outsource control centre operation to another country;
2063 – Track-to-train communications have eliminated intermediate signalling. All trains are intelligent with each other and no signallers are now required. Maglev technology has matured such that over and under passing points permit conflicting moves safely.
Fanciful? Well maybe some of it is, but good that the budding engineers of the future can create visions of how things might evolve.
Doug McCormick of Atkins brought some realism to the proceedings. He believes that it is paramount that the industry does not lose sight of safety yet still drives innovation forward. However, the global shortage of signal engineers will prevent technical advancement unless more people can be attracted into the industry. This in turn must not compromise standards and quality and the focus needs to be on the future, not on the past. Big thinking is needed but taking recognition from the solid foundation of signalling’s heritage.
So, the day gave a fascinating insight as to how S&T engineering has evolved from the past. Some predict that the biggest changes to the profession will happen in the next ten years. Only time will tell.